Loading summary
Transcript49 lines
- [00:01]
A
I have some really exciting news. We just released our second annual AI Marketing Industry Report and it's free. You'll be surprised by how often marketers are using AI and the tools that they're using. This is a comprehensive study of more than 730 marketers, and it covers how marketers are applying AI to their work, the benefits of AI, the big concerns marketers have, and a whole lot more. Even though I'm the author of this research study, I can say with full confidence that this is the most comprehensive study of AI adoption among marketers that I have ever seen. Get your free copy now by visiting social mediaexaminer.com airport25 this might be just what you need to get your boss or your clients moving along with those AI initiatives you want to start. Get it now@social mediaexaminer.com airport25 hey, it's Michael Stelzner here. After running Social Media marketing world for 12 years, I've noticed something fascinating about the marketers who attend. They usually fall into two camps. Camp number one is those who show up because they have to. Their results are declining. They're scrambling to catch up. And camp number two is those who show up because they want to. They're already achieving a level of success, but they know that change is coming and they want to go so much further. Both learn valuable insights, but their outcomes are completely different. As Emily Ray Shutti said, I came away with a million ideas and new creative ways to approach my work. My business has since doubled in size and revenue. The AI revolution is here. Instagram algorithms keep shifting, Facebook's ad costs keep rising, and the list goes on and on and on. What type of marketer will you be in 2026? Join thousands of marketers at Social Media Marketing World 2026 this April in Anaheim, California. Save big when you register today at Social MediaMarketingWorld.info.
- [02:02]
B
Welcome to the AI Explored podcast, helping you put AI to work. And now, here's your host, Michael Stelzner.
- [02:14]
A
Hello, hello, hello. Thank you so much for joining me for the AI Explored podcast brought to you by by Social Media Examiner. I'm your host, Michael Stelzner, and this is the podcast for marketers, creators and business owners who want to know how to put AI to work. Getting AI to work in business is difficult. People say that your competitors who adopt AI are a threat. But is that true? In today's episode of the AI Explored Podcast, we'll explore how traditional businesses can successfully put AI to work. My special guest is an AI strategist he's co founder of Xero to 60 AI spelled out 0260 AI a consultancy that helps mid sized traditional businesses successfully implement AI. He's co host of the AI accelerator for Business show. Carl Yeh, welcome to the show. How you doing today?
- [03:08]
B
Good, good. Thanks a lot Michael. I've been a long time listener of well especially on Social Media examiner which you know, you're really well known for. So yeah, just a pleasure to be here.
- [03:18]
A
It's awesome to have you here. So let me start with this first question. How in the world did you get into AI? Cause I know you're like most of us, there was a journey, right? So so like you were known for something else, now you're getting known for AI.
- [03:30]
B
It actually started in 2014 where I think you and I probably met 2013, 2014 at content marketing World. And it was actually, I think most of your audience would know a person named Paul Raitzer. I was actually at one of his sessions and he had a book called I think the Marketing Blueprint and I actually asked him to sign it and he was signing it and I just have a quick conversation with him and he was talking about all the stuff that he was doing in terms of AI or getting into. So that's kind of the same time I started looking. But really where it got my really piqued My interest was 2018 at Google IO. I wasn't there. I remember where I was at the time and I was watching Sundar do this restaurant reservation through AI and I thought this is the coolest thing I've ever seen. And this is where I really dug deep because I thought hey, this is coming now. Well obviously it didn't 2018, 2019, 2020. Played with the tools, you know, GPT2, GPT3, some of the image. And I obviously no one knew that anything would be even the adoption would be this high in terms of people adopting it obviously until November 2022. And that's really where everything took off and coming from that content creator, marketing background, just started creating content on it and some businesses saw it that I used to work with, wanted me to come in and speak and then kind of snowballed around that. Started learning because I have a bit of a background in coding, started learning more of like you know, building retrieval, augmented generation, things like that. And then my business partner kind of jumped the gun and incorporated us in I think September 2023 and here we are.
- [05:20]
A
I love it. And you have been working with a lot of traditional businesses and what I mean by traditional businesses is businesses that have been around for a very long time. And we're going to get into some of this in a little bit. But what are some of the biggest misconceptions that you see when it comes to AI adoption? Because obviously you're talking to a lot of people, you're hearing a lot of things.
- [05:39]
B
Well, I think that really for me, I three see three major areas. The first one is the assumption that it's easy. And when we talk about adoption, we're not talking just adoption for one month. We're talking you have to have sustained adoption, right? 3 months, 6 months, 12 months. Because there's going to be some money and some resources put into it by the business. And also the assumption that it's just bolting AI onto your current processes and workflows, which if you do, yes, you'll get some productivity gains and efficiency gains. But I think doing this for two years now, we're over two years, we're starting to notice. And this is kind of the thought I had was bolting on AI to your process, your workflows or even a product isn't good enough. There's also there was that old story about steam engine, right, where it wasn't until the the system was built for electricity, not for steam, not where you just replace the steam, an electricity with a steam engine, but you have to actually build the entire, rebuilt the entire process for electricity that the productivity gains were there. So what we're looking at is like you have to rebuild your processes and systems from ground up, being AI centric, not just adding AI to it. I think the third big piece is, and I think the most important piece is the change management piece, right? To do AI adoption, you have to change people's behaviors because they've been trained, whatever profession they have, years, decades, as well as each company has built their processes over many years, especially traditional companies, where it's years or decades of process, of over process, of complexity with more process. Right. So bolting on AI to that is very, very difficult. And we'll go through some examples later on. But that's, that's really why from AI adoption misconception, these are the three things that we're always noticing.
- [07:47]
A
What's your take on the phrase that people say all the time, which is you need to get ahead of your competition or they're going to crush you because they're embracing AI? What's your view on that?
- [07:57]
B
Yeah, your competition will use AI and maybe they'll get a little bit of head. But really you know, the, the danger that you should be looking for is not really your competition using AI. It's native AI companies, right, that don't have change management, don't have to bolt on processes. All their processes are already AI centric. Their people are already very knowledgeable or trained or very flexible in adopting AI into their workflow. And I think one of the most important things is this has nothing to do with current business operations because this is how businesses have been built. But there's a lot of business processes and operations of legacy and current businesses that just don't jive with AI, right? So, for example, like the way it rolls out technology, whether it's, and then let's say annual budgeting, long RFP processes, waterfall project management, some of these things don't jive. Where AI is, you know, the progress is within weeks, within months. And your business has to be ready to adopt or be flexible enough to change with the new tech because you're, you know, the tech could be cannibalized by something else or the major AI labs could release a feature that could wipe out what you've been using. And it's now $20 a month instead of you paying 30, $40,000 a month.
- [09:32]
A
It's a really fascinating concept to think about the fact that, yeah, your competitors who embrace AI, your known competitors who embrace AI, there's this race that you're all in, right? And maybe they'll be a little bit ahead of you one week and you'll be a little bit ahead of them the next week. But what I'm hearing you say is the bigger opportunity or threat, if you will, is going to be these startups that are going to come completely without any preconceived notions of how things ought to be done. They're going to be very AI focused, they're going to leverage AI to give them an unfair, you know, what we say is unfair, but, you know, a competitive edge. And they're going to come and they're all, all of a sudden going to just rush ahead of you. It's kind of like you're in the Titanic and they're in a little speedbo, but, but their little speedboat is just as powerful as your Titanic. That's really the threat that we're facing here, which is interesting, which is why I think it's so important we talk about getting implementation of AI right inside the business. So what I want to ask you is why is this so important for businesses today, especially traditional businesses? Said another way, if people listening, pay attention and Implement the things we're going to unravel today when this is done. Well, what is the upside here?
- [10:44]
B
Obviously right away you'll, you're going to get your productivity efficiency gains, right? So let's say you do bolt on AI into your current process and workflows, you'll get that productivity gain. And we're not saying in 95% full autonomous workflow, 100%. No, we're. Our big thing was always been 10 to 40% to start and then kind of build on from there.
- [11:10]
A
10 to 40%.
- [11:11]
B
What improvements, productivity gain, efficiency gain, improvements to your system, whether it's 10. And I think there's a little bit of management, you have to manage expectations because some, what we're noticing is some businesses and some people think, oh, AI can do all this stuff for me, probably because of the hype cycle. But then also let's say, you know, we implement an AI, whatever AI solution, automation, whatever it is, and let's say it doesn't work a hundred percent. It's like, oh, you know what? Not for me, it doesn't do my or that task 100%. So I'm going to keep doing what I used to do. It's like, well, no, it can do 30%, 40% better. So why are you going back to what you were doing rather than, hey, we're 30% of the way, you just cut out say seven hours a month, that's pretty good, rather than continuing to do, you know, 20, 30 hours. So that's what I mean by immediate product efficiency gains. But also I think it's more important when you're training or retraining your own workforce, you're helping kind of the rising tide lifts all boats where if everyone in your company is just 20% better, let's just say at prompting, 20% better, at prompting, you're going to get that productivity and efficiency gain without any type of custom solution or built automation, whatever them just having that knowledge. And it's also preparing your workforce for this new AI economy that's coming and starting to inject kind of a new way of thinking. Like I was saying, with an AI native company being flexible so when new models come in, they can adjust, right? Different types of tools. And back to the issue about, you know, AI native companies coming and, you know, being competitive in your industry. This is what you need to do is you need to train your team, get a new mindset and focus because that's what you're competing against. And how do you compete against a company or people that are Native. Well, that's the level up the skill set. And then, you know, you obviously have to keep your business going. So this is where your productivity and efficiency gains happen. But you also have to look at the long term too is, you know, your employees, the mindset of what your, where your company is going to go in the next five years and where AI is moving as well.
- [13:49]
A
Perfect. Okay, so we are here to talk about how to successfully deploy AI in a business that is been around for a while, right? Some sort of a business that's a traditional business. But anyone who's listening, who works for a business that's been around for more than a decade, you know, might be considered traditional. Even my business that's been around since 09 might be perceived to be traditional in my niche because obviously there's been a lot of change. So where do we begin if we want to implement AI across our business, where do we start?
- [14:17]
B
Usually you'd start with the discovery, right? So when you talk about discovery, you're looking at all the processes and workflows that your company already has, specifically things that are repetitive, things that are data driven, things that are predictive. This is, I took that from, from Paul Raids course and really what we call the eye roll tasks. So if your employees come in and whatever tasks they roll their eyes at, look at those things. And when I talk about discovery, you don't actually talk very much about AI. You just learn about how those workflows and processes operate. And this is good whether you implement AI or not, because most businesses we've worked with haven't done this kind of audit of their own processes, right? So this is where you begin. You sit with your departments, with your teams, take an hour, two hours, maybe even three hours, and just list those processes, list how they work, all the, the individual operations, and sometimes it turns into a complaint session. But that's okay because that's what you want to do is you want to get that feedback not just from your leaders and executives, but from those frontline teams that are dealing with this day in, day out, day in, day out. And that's where you usually start. So for example, I we're currently doing a discovery where we just finished a discovery with a land development company, right? And you know, there's a whole list of processes, but two very interesting ones that just happened. The first one is there is somebody, and again back to decades, layers of complexity, that one of their tasks is to go into a folder, look at I would say 50 subfolders, and find if a specific file is there. So that person just goes into every single folder to check if a certain file has been generated by their system. So that's one. Right. And that sometimes it takes that person an hour, two hours, three hours to get through that list or looking for different things. Another task would be another person has to generate five to six business intelligence reports every Monday and extracted from a whole set of PDFs that are updated throughout the week. Right. Just two simple tasks that. And I'm sure most people are starting to think, hey, wait a minute, AI can probably do those things much faster or augment it to be not a hundred percent, but maybe 10 to 40%. Just help that person who's. It's a data driven. It's a repetitive task.
- [17:06]
A
Okay, this is really interesting. Obviously the bigger the company, the more there's a lot of these kind of things. Right? So like if you're dealing with a company like in my case that has like maybe you know, 20 people versus a company that has a hundred or a couple hundred people, I guess the real question is like, first of all, let me back up the train. What I'm hearing you say with this eye roll task is we want to find the things that people do that frankly they just wish they didn't have to do. Right. They take time. They involve finding information, they involve doing specific things over and over again, as you called repetitive predictive data driven tasks. And these things happen all the time. Sometimes they're updating spreadsheets. Right. Other times they're creating PDF', other times they're, they're notifying people that a folder has been, a file has been dropped into a folder. All these kinds of things, right. So there's so many of these things that happen inside of our business, Carl, that it's kind of like, do you start with one department typically, do you start with one person? Because this could be like a. You could come up with a list of a hundred things really quickly.
- [18:13]
B
Yeah, yeah. So each department, we usually end up with seven to eight major things that they identify. That's like kind of a more of make work tasks. You know, it's important that work is important in the overall. But it's sort of like, why am I doing this? Like, why do I. Like exactly the eye roll. So each department, some company is bigger than others, smaller, but usually four to seven departments that we usually deal with, but depends on, you know, your business. The thing is, before you even start discovery, two, some things you have to think about is. And you have to think about the change Management pieces. Is the people in this department or team ready for this change? Because that's a big piece. Two, obviously data cleanliness. How clean is their data? Because sometimes you're getting data sets all over the place and even with great intentions, there's still a lot of data cleanup that you have to do and maybe that's a priority as well. So there's those kind of things you have to look at, but from a discoveries perspective at the very least. Now you've outlined all these different eye roll tasks and sometimes we end up with 40 to 50 per company that have been identified. Right. And it's a lot that you can get into.
- [19:38]
A
Well, I love this. For anybody who works inside of a decent sized company and you want to try to deploy this strategy that Carl's talking about, I can see how the rationale logically makes a lot of sense. Hey boss or boss's boss. I believe that I can save the company substantial amount of money, possibly freeing 10 to 40% of the time of the people that do some of these kinds of tasks. Would you be willing to authorize me to do an audit of opportunities to use AI to basically speed up the amount of time that it takes to. And the first step of the process is for me to identify early targets. Right. Where AI makes a lot of sense. And I would imagine they would probably say, sure, right. Because what person who either is a division head or runs a business doesn't understand that there's an enormous amount of process opportunity improvement, you know, inside the business. So once we've done this discovery phase, what comes next?
- [20:29]
B
So you take the notes and usually you, I highly recommend you record these sessions just because there's going to be a lot and it's a lot to take notes of. You take those and then you actually develop a discovery summary for each department or team that you did with. Right. So you have key themes from that discovery session. This is where you look at the workflows and processes identified and then connect it to AI solutions. Right. So, and we can talk about solutions later. But like, so you have those lists of solutions that directly address those workflows and processes, the eye roll tasks and so on. But you also can now identify non AI solutions because there's a lot of times when you're looking at some of these processes and you're like, wait a minute, this could be something completely different. Like for example, we were working with a, I don't even know how you describe a company that develops infrastructure for dentists and, and hospital offices. You see the little children's, I think, play areas and whatnot. So working with them. And it was interesting where one department said, hey, you know, we spend two, three hours a week re updating the format of a file that one department sends to us. And we're like, wait a second. So we just walked over to the, that department. Hey, could you output this in this format? Oh yeah, no problems. That is not an AI solution. That is a pure communication issue. And that's been going on for, for years. One department just didn't talk to the other or made assumptions. So you'll find a lot of these too. And so you can, you know, you can address a lot of things.
- [22:16]
A
Have you ever attended a conference and came back with all sorts of ideas and you were fired up but you struggle to get everyone else on your team to be on the same page? Well, here's what I noticed after a dozen years of social media marketing world. When one person attends, they get great ideas. But when a team attends together, they create something much bigger. Think about it. One person learns about AI and marketing. A team builds out an AI strategy. One person discovers Instagram tactics. A team redesigns their entire content approach. Noah Stanley told us, quote, I will bring multiple members of my staff next year. The mindset, openness and amount of time the experts gave to us was head and shoulders above other conferences, unquote. When your team attends together, you can cross pollinate ideas in real time, divide and conquer to cover more ground and return with a unified action plan, not just a bunch of scattered notes that you never get around to. Plus, you've got a built in support system when you try new strategies and you're back at the office. The company seeing the biggest results from social media marketing world, they're the ones investing in team transformation. Bring your team to Anaheim this April and multiply your impact. Learn more at socialmediamarketingworld.info Change Management is a phrase that I'm hearing over and over from a lot of guests on the show here. And managing change inside of a structure, a system, an organization, does require human involvement. And I like the fact that you're saying here, hey, some of these things need to be fixed because people are actually not putting quality together in the first place, right? Or they're not putting it in a format that the recipient can use it. And if we can solve for that human to human, eventually that also might mean that could be optimized down the road with AI. Is that what I'm hearing you say?
- [24:09]
B
Yeah, yeah. And again, there's a lot of things that companies do, years, decades of process that no one's actually reviewed. There's definitely AI solutions, but then there's non AI solutions as well. So building on that, there's another section where hey, what are the process and workflows? You have to identify these that we can completely rebuild to be AI centric. Now like could we knock off this entire process and just rebuild it? And this is where you have to start. This is part of the thinking of hey, you know, we've been doing this for so long and now this is kind of the mind shift your people will need. So you'll have that and then you have a roadmap, right? How? Six months, 18 months plus and so on that roadmap for that specific department or team. So there's going to be these discovery summaries for every single department with all these sections. And then based on all that conversation, all these discovery summers, you can create an AI strategy for your entire company that has everything that you went through, all the pilots. But then in that strategy you can pick five to seven that your company can start with. So of the 40 to 45, let's pick five to seven, three to four high priority, that you can put resources and people behind and focus and then pick 1, 2, 3 processes that you want to rebuild entirely to be AI centric. So that's usually, you know, when you have that strategy, you have those pilot recommendations, rebuild recommendations and a roadmap. Your company is, you know, you're well positioned, you've connected the problems with solutions. You've got input from Frontline. You know, you have like an overall plan for your company, not some hodgepodge of here's a group using AI, here's another group, here's one group. Maybe you have an AI council, but it's kind of disjointed. You have something your business can work towards.
- [26:18]
A
You have an example. Yes, let's hear it.
- [26:20]
B
For example, we worked with an accounting company and we did this entire strategy discovery summary. I think they had six departments and we full, full on strategy and you know, we recommended six AI strategies. They accepted four of them to get started. So for example, one of them is an overall of their entire manual data entry. So instead of them, you know, t slips bank PDFs that people someone was manually entering into a spreadsheet which then would connect to QuickBooks. That was built to be an N8N automation from document recognition. We used Mistral's PDF reader. There's an actual, there's a template version of that and then connect it to their spreadsheet, actually their CSV and then connect it to QuickBooks Online. So that was a connection for them. And then they also had a AI phone reception. It was pretty much using a third party voice reception mostly to handle like the quick repetitive calls and help them book consultations because there was, it was taking a lot of time for their reception to go through this. So those are just two examples of the pilots that they launched. And I think now we're calling into kind of the, the scale phase. But yeah, they accepted three or four of them. You know, they're moving on.
- [27:52]
A
Love it. I got to ask because people want me to ask. Okay, Mistral, for those that don't know, I believe that's a French based LLM, their PDF reader. Tell us a little bit about that and then also tell us about this AI phone reception third party tool that if you remember who it was. Because I want people to kind of just know that those things exist, you know.
- [28:10]
B
So for Mistral, I think them and Cohere, so Cohere is a Canadian company, so plug for, for Canadians, they have one of the best PDF ocr. Optical Character recognition. Yes, that's right. They have I to me and we've tested it with different large language models but they had the best one that could recognize a PDF in different formats because different vendors would or different businesses would submit different PDFs and be able to extract the information that we were looking for and then put it into the headlines, into their spreadsheet. Right. It did the best. It was, I would say 95, 97% reliable for were the most repetitive tasks. So that's an open source model. We built it on premise. So it's not cloud based, it's on their servers. I kind of, that's why I really like N8N because of that on premise solution. And then for the voice, AI voice, I mean AI reception, we just use bland, bland AI.
- [29:12]
A
So just for folks that don't know, N8N is a tool that's kind of like make or zapier, but you can host it on a server and then it's kind of like WordPress, it's kind of like open source, so you can lock it down. So what, when you said you built it on premise, you meant physically on site, not a tool called Premise. Right. So you built a local version of N8N on one of their servers in the building so that there was never any concern that any of this information was going to leave the building. Is that Effectively what I, what I'm hearing, you say yes. And then you also said that phone reception was called bland. Is that what you said?
- [29:49]
B
Yes. Blind AI. Yeah.
- [29:50]
A
Interesting. Okay, very cool. People love the tool. So, all right, so now let's move into the pilots. Like, give us some insights on, like, we've done some discovery at this point, we've presented to our boss or client in your case. Here are some things that we believe will have a material impact on your productivity. And so let's talk about how to actually do pilots. Like, any insights on that?
- [30:12]
B
So when you're looking to do pilots, what we've learned over the past two years, there has to be a dedicated think subject matter expert. Whether you use third party consultants like us or you go internal, you have to dedicate at least one, preferably two, as your backup subject matter expert for whatever the task is in whatever the department. So that person actually has to dedicate some of their time. It can't kind of be at the side of their desk. Like you have to dedicate that resource because there's a lot of that pilot probably won't get launched unless it's a simple prompting. But even in prompting there's a lot of back and forth that you need to go through. So for example, one of our clients as a construction company and their task was invoice consolidation. So they have to take vendor PDFs, so like Rona, Home Depot, like they get a lot of these PDFs from different vendors and they have to check if the invoices on the vendor statement matches the invoices in their system. So literally there is one person having to manually look at each invoice and into their system by code to see if the numbers match. Right. Well, this is a pretty simple thing with AI, right? So first, you know, we obviously tested it with, you know, comprehensive prompt, just drag and drop the PDF, drag and drop the CSV of all the statements for that month because they have to do it pretty much weekly and then have it review, look at any discrepancies, which ones are there, which ones aren't and how much. Right. But that takes like a good one, two weeks. It's not just for someone who is AI native and can do this. Yeah, you can probably do it. But think about the person who doesn't use AI or barely uses AI or even just uses AI like as you would Google, to be able to develop that prompt, adjust the prompt, and then test it over and over and over again to ensure that it's reliable because that's what you want to do, right? You don't want to be looking at the results and then rechecking. That defeats the whole point. You want to make sure it's reliable and, you know, go forward from there. So we did that with ChatGPT3, ChatGPT5 made it even better. But this is where, when you're looking at pilots, you have the pilot phase, but you also want to scale it real quick.
- [32:56]
A
Before we get into scaling, we're going to get to scaling in a minute. But I want to get back to the internal champion. Give me some guidance, like on how you give your guidance, like to, hey, Company X, I need a person internally to be the champion. Like, what does that mean? How do you convey them? What is the expectation of the internal champion?
- [33:14]
B
So there's a difference between the subject matter expert and the champion, right? So the subject matter expert is someone who knows about the topic, doesn't necessarily have to be 100% into AI.
- [33:24]
A
The internal subject matter expert on the process.
- [33:27]
B
On the process, yes.
- [33:28]
A
Okay.
- [33:29]
B
The AI champion, though, these are the people, ideally in the company who already use AI, who are excited about it. Probably bugging leadership to move forward a little faster. A lot of the times gets pushed down because, you know, the company's like, well, we're going to go there, but you're using it, just use it yourself and so on. So it's that person and what you want to do with that person, and ideally a group of those people is form yourself a committee. And the reason why you need that, and not necessarily to develop policies and governance, but it's more. So the two years we've been doing this, with the hundreds of businesses that we've been meeting and our clients, is that back to the whole sustained AI adoption, Because you need a group to keep people accountable, to keep using whatever solution you had. So even if you build the best AI solution. So for example, this N8N solution for PDFs, right, for the accounting company, do you still need someone to put those PDFs into that folder? Like, that's all you need to do. But what we noticed is after the first month, they stopped putting PDFs into the folder to kick it off. And you're like, well, why do you stop? It's like, oh, I forgot, I got busy. I kind of just thought the process was a little cumbersome. Even though it's just a drag and drop. It's those kind of things, right? So. And you have to keep people using it accountable. And, you know, it's like Oh, I had questions about this, but I didn't want to bother you. It's like there's a lot of human behavior you have to change and there's a lot of habits you need to break. So that's why your group of champions has to meet every month and talk with people, say, hey, are you using it? That's the number one question. Using it? Why aren't you using it? Are there any issues with it that you know, didn't come up? Like, those are the things you have to do. Because we have to remember outside of the AI bubble, outside of X, you know, all the people are excited. Most people are like, oh, okay, AI is cool, it's part of a tool. But like I have stuff to do at work, I have stuff to do at home, I have family, I have all these different things that AI isn't top of mind. AI isn't. It's not my daily process. And we have to get people to think about it more and use it more to get that sustained AI adoption.
- [36:05]
A
Does this mean, Carl, that it's really important to get buy in from kind of the top or near the top of the organization, like the division head and to just like have them say, hey, look, this isn't just like a fun little thing we're building for you and you can use it or you can't use it, but to mandate change or do you feel like bottom up approach the better way to go? Like, what's your thoughts on this?
- [36:26]
B
It's a fine line because there's a danger with either. There's a danger if you have a mandated session, mandated from top down. Hey, you have to use AI. But we've run across that where they've mandated the company. You got to learn AI, you know, use AI. But there's no connection to. This is why the discovery is important. There's no connection to actually helping this person use AI. Worse is if they actually leaders get excited about a demo they've seen before, buy a tool and then enforce that tool into the company. An example would be Microsoft Copilot. Copilot's good for some things, but buying 200 licenses. And then we get called and hey, could you retrain our staff? Because we only getting 10 to 15% usage of this. It's like, well, why did you buy this in the first place? And it's like, oh, we thought it would solve this problem. It's like, no, that's why you need to talk to your people. Because that is not the problem. And this tool, it doesn't jive with what they're looking for. It's not the solution for what their issues are. So the top down is important, but also the bottom up. You can't just have like, oh, groups of people using AI. There has to be some sort of coordinated effort because again, rising tide lifts all boats. So everybody has to be part of it to get the full business benefits. Because it isn't. Again, it's not just your competitors that you're worried out for. It's that AI native company that's coming that's like, oh, I don't have to. There is none of this problem. You know, they're starting off with a massive head start.
- [38:02]
A
I really like this conversation. I think that since we're in pilot mode right now, the people that are part of the pilot need to understand what's at stakeholders. We need to explain to them, look, first of all, what's the benefits to them, right? If you employ this, those eye roll tasks will dissipate, they will become less necessary and with that newfound time, you can work on higher order tasks like analyzing things or being creative or working on strategy and then. But the other side of it is you also need to understand that we as a company are going to eventually have newer businesses that are AI savvy, that don't operate off these old scripts that we've been using for decades that are going to come along and they're going to crush us. So there's the flip side of it, the problem. We have to embrace this because we have to understand that the world is going through a paradigm shift and we can either be survivors or we can be effectively knocked off. Right? And that is an important discussion that needs to happen. So they need to understand what's in it for them, but they also need to intimately understand what the why is. And then they need to be reminded of that, I would imagine, right? Like, hey, this is something that's happening and even though it doesn't feel like it's happening, it's kind of like a storm alert. You know, there's a storm coming. Oh, they never come. There's a storm coming. Oh, and then the storm. Ah, there's a storm. Yeah, yeah, that's kind of, that's kind of like. So what about ongoing training and stuff? Like, I mean, do you recommend that? And because obviously when you're in the pilots, what you thought in reality might not match, right? There might need to be some changes.
- [39:43]
B
So when we've seen business training, right, you have, you could have some sort of learning Management system and you take the courses there or you have a lunch and learn. But with AI you kind of need to have it on a regular basis. So even just having one training of basics of AI, let's say prompting or whatever, because of all the changes, even the training we did for a company three months ago, we kind of need to update it because for example, a majority of the people in that company are using ChatGPT. Well with GPT5 you have to adjust how you prompt because there's just different the way that it operates and it over delivers. Sometimes you have to go back and retrain that. So that's something a little different for a lot of companies because it's usually, Hey, 1, 2 trainings were done, it's like, no, no, you need to keep going. And this is part of going back to your AI committee or AI champions. It's not just, you know, reviewing projects, it's like, hey, we gotta develop these training systems ongoing to let people know, hey, you gotta change it a little bit because this new feature, this new tool, this new way of doing things, you have to adjust, if you step back a bit, that's a whole different way for businesses to operate where you have to constantly change not just the tool feature but how you operate. That's a completely different way for businesses to operate. And that's what I mean, right? So because the training goes hand in hand with the mindset and the change management that your company needs to go through to do that. And if we're talking about, you know, where is the storm coming from, where is this, you know, where are these AI natives? All you have to do is go to Y Combinator, go for the request for startups, their summer 2025. They literally say it on the thing. Instead of working with the dinosaurs, you can make them extinct. Like they're looking for startups to go native against all these traditional industries, so they're funding it. There will be companies in your industry that are going to be a native that will maybe have significantly less people that can do significantly more.
- [42:01]
A
Okay, that's kind of crazy, but scary, but also like nothing like a little fire under the bottom side here. So like, okay, we've, let's say we've successfully had a pilot. So what's next?
- [42:12]
B
So what's next is how can you scale that pilot but also have to remember your pilot has to be flexible enough to change, right? Because the tools are changing, the techniques are changing. So an example here would be going back to the construction company, the invoice, consolidation we initially used ChatGPT03 to do that consolidation. You know, cut a lot of hours out of them, but they still have to drag and drop that. So the next thing that we did is a custom GPT. Made it even easier. You know, there's a custom GPT for you, similar instructions, drag and drop. Cool. The next thing, though, is because ChatGPT has connectors. You connect it directly to the file. So you don't even need to drag and drop anymore. You just, hey, in this file, here's the prompt execute, and away it goes. So that's part. That's the next part. Now we're looking at, hey, with ChatGPT's agent mode we tested, hey, can the agent do the same thing we were doing, copy and paste manually? Yes, it can. Combine that with connectors and combine that with their scheduled task. So now there's no dragging and dropping. There's no even opening the task. You have a scheduled agent to do this invoice consolidation every Friday. So now this agent executes every Friday. This consolidation, you know, you do a spot check here and there to make sure. But that's kind of the next piece that we're testing out to ensure reliability. But that's kind of an example of how, you know, your invoice consolidation not just scales, but improves over time with the tools that are coming with it.
- [44:11]
A
Wow, Carl. I mean, this is crazy stuff that we're talking about, right? Like, who would have thunk a couple years ago we'd be talking about what we're talking about today? But it's really, really cool. And I know we've simply just scratched the surface of what you and your business do. If people want to connect with you, Karl Ye on the socials, do you have a preferred place platform? And if they want to work with your business, where would you like to send them?
- [44:33]
B
Three major things. One is zero to 60 AI. That's our website. Two, you know, we run a show called the AI Accelerator for Business Show. YouTube, your preferred podcast platform of choice. And finally, you can find me on TikTok. I do a lot of videos there on everything that we talked about and more.
- [44:52]
A
Thank you, Carl, so much for sharing your wisdom with us today.
- [44:55]
B
Thanks, Michael.
- [44:56]
A
Hey, if you missed anything, we took all the notes for you over@socialmediaexaminer.com a74. Be sure to follow this show on your favorite podcast app. And if you've been a longtime listener, would you give us a review on whatever platform you're on and let your friends know about this show and do check out our other shows, the Social Media Marketing Podcast and the Social Media Marketing Talk Show. This brings us to the end of the AI Explored Podcast. I'm your host, Michael Stelzner. I'll be back with you next week. I hope you make the best out of your day and may AI help you become more successful.
- [45:29]
B
The AI Explored Podcast is a production of Social Media Examiner.
- [45:36]
A
This is the year to finally come to Social Media Marketing World 2026. Grab your tickets right now by visiting social mediamarketingworld.info.